REVIEWS MEMORY SYSTEMS The anatomy of memory: an interactive overview of the parahippocampal– hippocampal network N. M. van Strien*||, N. L. M. Cappaert‡|| and M. P. Witter*§ Abstract | Converging evidence suggests that each parahippocampal and hippocampal subregion contributes uniquely to the encoding, consolidation and retrieval of declarative memories, but their precise roles remain elusive. Current functional thinking does not fully incorporate the intricately connected networks that link these subregions, owing to their organizational complexity; however, such detailed anatomical knowledge is of pivotal importance for comprehending the unique functional contribution of each subregion. We have therefore developed an interactive diagram with the aim to display all of the currently known anatomical connections of the rat parahippocampal–hippocampal network. In this Review, we integrate the existing anatomical knowledge into a concise description of this network and discuss the functional implications of some relatively underexposed connections. In the more than 100 years since the first explorations of underexposed connections tend to be erased from the the parahippocampal–hippocampal network by Ramon common scientific memory. For this Review, we have y Cajal1, numerous detailed anatomical tract-tracing assembled the extensive anatomical PHR–HF connec- analyses (BOX 1) have been published. These studies were tivity literature, focusing on all known connections of sparked by the discovery of a prominent relationship one frequently used experimental animal: the rat. We between declarative memory and structures in the human introduce a new approach to describe the network con- medial temporal lobe, in particular the hippocampal for- nectivity that uses an interactive diagram to display the mation (HF)2; the importance of the parahippocampal complete PHR–HF connectivity (see Supplementary region (PHR) for memory was established only later3. An information S1 (figure) and Supplementary informa- *Department of Anatomy and increasingly complex picture of the connectivity within tion S2 (box)). The complex and detailed connectivity Neurosciences, VU University and between the HF and the PHR has emerged over the patterns in this diagram are made accessible through Medical Center, P.O. BOX 7057, 1007 MB years, and comprehensive knowledge of the PHR–HF the ability to switch on and off individual or groups of 4 Amsterdam, The Netherlands. network lies at the basis of understanding its functions . network connections between cortical layers and/or ‡SILS Center for The level of anatomical detail at which an experi- anatomical areas. The information this diagram pro- Neuroscience, University of ment must be carried out or results interpreted vides could prove to be useful at a time when research Amsterdam, 1098 SM depends on the questions under investigation. In is moving beyond the functional explanations that can Amsterdam, The Netherlands. §Kavli Institute for Systems some instances, the effects of experimental manipula- be provided by a PHR–HF circuitry model that contains Neuroscience, and Centre for tions can be interpreted using connectivity data at an only a subset of the connections; moreover, it might the Biology of Memory, overall network level (without taking the details of lead to a re-evaluation of the functional importance of Department of Neuroscience, local networks into account). Other studies require connections that have previously been ignored. Norwegian University of Science and Technology, more detail, but even those studies that benefit from This Review first describes the anatomical concepts N-7489 Trondheim, Norway. a detailed understanding of the circuitry often do not, that are essential to understanding the PHR–HF cir- ||These authors contributed for a variety of reasons, take all the known connections cuitry (for an extensive description, see REFS 5–7). Next, equally to this work. into consideration. Sometimes connections are sim- it presents an overview of the main PHR–HF circuits as Correspondence to N.M.v.S. ply overlooked, whereas other times connections are well as of some of the lesser-known aspects of the cir- e-mail: n.vanstrien@temporal-lobe. intentionally left out because they seem to have no cuitry, using the interactive diagram (Supplementary com function and are therefore considered irrelevant for a information S1 (figure)). Subsequently, it shows how doi:10.1038/nrn2614 particular theoretical interpretation. Eventually, such having detailed knowledge of the PHR–HF circuitry can 272 | APRIL 2009 | VOLUME 10 www.nature.com/reviews/neuro © 2009 Macmillan Publishers Limited. All rights reserved REVIEWS Box 1 | Neuroanatomical tract-tracing methods consisting of medial (MEA) and lateral (LEA) areas), the perirhinal cortex (PER, consisting of Brodmann Most of what is known today about the pathways that connect neurons in different areas (A) 35 and 36) and the postrhinal cortex (POR). brain regions has been discovered by using neuroanatomical tract-tracing The PHR is generally described as having six layers. The techniques156. A tracer is a substance that allows such pathways to be visualized. coordinate systems that define position within the HF Tracers can be injected intracellularly to label the dendrites and axons of a neuron. FIG. 1 Both autofluorescent dyes (for example, Lucifer yellow and Alexa dyes) and and the PHR are explained in . biotin-derived dyes are often used for intracellular labelling, as they can be easily visualized using fluorescent microscopy. Alternatively, a tracer can be injected at a Circuitry of the PHR–HF region stereotaxically defined extracellular location in the in vivo brain. The tracer is taken up In the interactive diagram (FIG. 2; Supplementary infor- by neurons at the injection site and is transported or diffuses within cells. A tracer mation S1 (figure)) we attempted to display all of the substance can be transported anterogradely from the soma towards the axon PHR–HF connections that have been reported in the ana- terminals (for example, Phaseolus vulgaris leucoagglutinin), retrogradely from the tomical literature concerning the rat (for references see axon terminals towards the soma (for example, Fast Blue), or it can be transported in Supplementary information S3 (table)). The interactive both directions (for example, horseradish peroxidase). Another tract-tracing method diagram contains almost 1,600 connections, which can be involves creating small lesions and visualizing the resulting degeneration; the labelled displayed at a customizable level of complexity. This allows connections are generally assessed using light microscopy. Electron microscopy can be used to visualize whether a presynaptic axon contacts a postsynaptic element. This is a easy comparisons between the detailed PHR–HF circuitry very accurate but time-consuming method because only small pieces of tissue can be illustrated by the diagram and a ‘standard’ model of this examined at one time. Alternatively, confocal microscopy allows three-dimensional circuitry (FIG. 3), which displays the subset of connections reconstruction of larger pieces of tissue and can indicate whether pre- and that are currently most often used in the field (based on an postsynaptic elements are likely to form a synapse. A question of current interest is analysis of a selection of recent key studies8–15). whether confocal microscopy is reliable enough for indicating such contacts. In order to increase our understanding of the connectivity of the brain and its related function, Connectivity within the PHR. In the standard model accurate numbers that provide information about pathways’ projection intensity and (FIG. 3), the projections from the PER and the POR to the 157 termination density are needed. To achieve this, techniques using viral tracers EC are often depicted with a topology that emphasizes and new genetic tools158 are being developed. PER-to-LEA and POR-to-MEA relationships. However, as can be seen in the interactive diagram, the available data indicate (see figure 1a in Supplementary informa- aid one’s understanding of some of the functional pro- tion S4 (figure)) that the POR also projects to the LEA, cesses that engage the PHR–HF regions, such as memory although quantitatively to a lesser extent than the PER formation, spatial navigation and temporal dynamics. (4.9% versus 15.6%, respectively, of the total cortical input)16. Likewise, the PER also projects to the MEA (see Hippocampal–parahippocampal anatomy figure 1b in Supplementary information S4 (figure)), con- The rat HF is a C-shaped structure that is situated in tributing a level of cortical input equal to that of the POR the caudal part of the brain. Three distinct subregions (7.5%)16. Neurons in layers II, III, V and VI of A35 and can be distinguished (FIG. 1): the dentate gyrus (DG), A36 of the PER project in a convergent way to LEA layers the hippocampus proper (consisting of CA3, CA2 and II and III17, whereas the PER projection to the MEA arises CA1) and the subiculum. The cortex that forms the mainly from A36 (REFS 16,17). The POR projection to the HF has a three-layered appearance. The first layer is a LEA arises from layers II, III, V and VI and terminates deep layer, comprising a mixture of afferent and efferent in layers II and III16,17. The POR projection to the MEA fibres and interneurons. In the DG this layer is called originates from the same layers and terminates preferen- the hilus, whereas in the CA regions it is referred to tially